role of diffusion-weighted magnetic resonance imaging in

Med. J. Cairo Univ., Vol. 87, No. 8, December: 5143-5156, 2019

www.medicaljournalofcairouniversity.net

Role of Diffusion-Weighted Magnetic Resonance Imaging in

Discrimination of Endometrial Carcinoma from Benign Diseases

Affecting the Endometrium

NERMIN Y. SOLIMAN, M.D.*; MAHA M. ZAKI BILAL, M.D.*; MOUSTAFA M. MOHAMED AL-KHIARY, M.D.**; EMAN M.H.M. ABO EL-ATA, M.D.* and MARIAM Y.M. FARID, M.Sc.*

The Departments of Radiology* and Gynecology & Obstetrics**, Faculty of Medicine, Mansoura University, Mansoura, Egypt

Abstract

Background: Evaluation of the role of diffusion-weighted

imaging (DWI) in differentiation between endometrial carci-noma and benign diseases affecting the endometrium.

Aim of Study: To assess the efficacy and the diagnostic accuracy of DWI in discrimination of endometrial carcinoma

from benign endometrial and subendometrial diseases with pathologic correlation.

Material and Methods: This study included 100 female patients with endometrial and subendometrial pathologies, their ages ranged from 20 to 85 years. Conventional T2- weighted, pre- and post contrast T1-weighted and DW-MRI

images with b values 0, 500, 1000 s/mm2 were obtained. The mean ADC values of all lesions were recorded. Results were compared, correlated with final histopathological diagnosis

reached by hysteroscopy, endometrial biopsy and surgical findings.

Results: Patients were divided according to the final

histopathological diagnosis into two categories (category I:

included 40 patients with endometrial carcinoma and category

II: included 60 patients with benign endometrial and suben-dometrial diseases). In DWI qualitative analysis; all studied

cases with endometrial carcinoma displayed high signal intensity, however; there was overlap in DWI SI as regard

benign diseases. The mean ADC value of endometrial cancer

was 0.73 ±0.10 x10 -3mm2/s which was significantly lower than that of benign diseases 1.36±0.32 x10-3 mm2/s (p<0.001). The cut off value for differentiation between endometrial

cancer and benign diseases was 0.915 X10 -3mm2/s with sensitivity, specificity and accuracy of 97.5%, 88.3% and 92.0% respectively.

Conclusion: DWI combined with ADC measurement provides valuable information for evaluation of suspicious endometrial and subendometrial pathologies and is able to

discriminate endometrial carcinoma from benign diseases

affecting the endometrium with high diagnostic accuracy.

Key Words: Diffusion-weighted – MRI – DWI – ADC.

Correspondence to: Dr. Nermin Y. Soliman, The Department of Radiology, Faculty of Medicine, Mansoura University, Mansoura, Egypt

Introduction

ENDOMETRIAL carcinoma is one of the most widely recognized malignancies of the female

genital tract, predominantly influencing postmen-opausal women presented with abnormal vaginal

bleeding, which is a non specific symptom since some benign pathologies affecting the endometrium

may cause bleeding [1] .

Radiologists considered endometrial pathologies to be a significant diagnostic challenge due to the

potentially similar imaging features of benign and

malignant endometrial lesions in morphological MRI [2,3] .

DWI is a recently widespread functional imag-ing method used to show tissue features dependent on the arbitrary dissemination movement of water

particles, thereby giving a tissue contrast unique

in relation to that appeared on conventional se-quences [4] . DWI reflects alterations in proton mobility due to the changes of tissue cell density,

cell membrane integrity and extracellular space

tortuosity due to pathological conditions [5] .

By performing DWI utilizing variable b-values, quantitative analysis, in particular, the estimation of apparent diffusion coefficient (ADC) values is available [5] . Several studies reported the value of estimating ADC in patients with endometrial car-cinoma in distinction between normal and cancer-ous tissue and in discriminating benign from ma-lignant uterine endometrial cavity pathologies [6] . Areas characterized by hyperintense signal at DWI

and low values of the ADC generally correspond

with foci of high cellularity represents a malignant

tissue [7] .

5143

http://www.medicaljournalofcairouniversity.net

5144 DW-MRI in Diagnosing Endometrial Cancer

So, the addition of DWI to conventional MRI

can increase the specificity of lesion characteriza-tion [8] , in order to obtain an accurate preoperative

diagnosis and to avoid aggressive or unneeded surgical management [9] .

Patients and Methods

This study included 100 female patients. Their

ages ranged from 20 to 85 years old. It was con-ducted at the Diagnostic Radiology Department of Mansoura University Hospital over the period from

July 2017 to July 2019.

Patients were referred from Obstetrics and

Gynecology Department of Mansoura University Hospital & Oncology Center, Mansoura University.

The study was approved by our institution's ethics committee, and all patients gave their in-formed consent before inclusion in the study.

Inclusion criteria:

• Patients with suspicious clinical findings as abnormal vaginal bleeding.

• Patients with suspicious sonographic findings as

abnormal thickened endometrium not correspond-ing to patient age and menstrual status, and endometrial cavity mass.

Exclusion criteria: • Patients with MRI incompatible devices. • Claustrophobic patients. • Patients with pathological results could not be

obtained.

All patients were subjected to the following:

I- Full history taking with a special emphasis on:

(Age, menstrual cycle, duration of menopause, history of replacement hormonal therapy, his-tory of medical disease or anticoagulant thera-py).

II- Clinical examination that was done by our colleagues in Obstetrics and Gynecology De-partment.

III- Routine laboratory investigation for all patients

for example (complete blood picture, liver and

kidney functions). IV- Pelvic ultrasound examination.

V- MRI examination: All MR examinations were performed using

1.5 MRI magnet (Philips-Enigma, the Netherlands) with a phased-array pelvic coil. Supine position

was applied to all cases throughout the examination with head pointing towards the magnet (head first).

1- The conventional MRI sequences performed included non contrast T1-W sequence (TR 400- 600ms, TE 10-14ms) axial section, T2-WIs fast

spin-echo (TR 3.000-6.000ms, TE 100-110ms) sagittal and oblique axial sections, Post contrast

T1-WIs after fat suppression aquired in at least

two planes after injection of Gadopentetate Dimeg-lumine (Magnevist) at a dose of 0.1ml/kg body weight, the injection flow rate was 2ml/s, and the contrast injection was followed by a 20-ml saline flush, Gadolinium-enhanced T1-WIs was per-formed after DWI utilizing the following parame-ters TR/TE 621/18ms, slice thickness 4-5mm; inter-slice gap 1-2mm; FOV 40 x 28cm and matrix was 256 x 196.

2- DWI MRI sequence was done before post

contrast T1-WIs, by acquiring conventional T2- WIs with addition of strong diffusion gradients

using a single-shot echo-planar imaging sequence with the following parameters TR/TE 2640/80ms,

slice thickness 5mm; interslice gap 1-2mm; FOV 40 x 28cm and matrix was 160 x 87. Motion-probing gradient pulses are placed in the three orthogonal planes to generate isotropic DW images

using a b factor of 0, 500, and 1000s/mm 2 . Fat suppression was used for background using a se-lective partial inversion recovery (SPIR) or short

T1 inversion recovery (STIR).

All images were transferred to a workstation (Advantage Windows version 4.2-07, GE Health-care) and the DWI sequence was post processed

with commercial software (FuncTool, GE Health-care) to obtain apparent diffusion coefficient

(ADC). At first, the conventional and post contrast

MRI sequences were evaluated then we correlate

them with DWI and ADC map.

A- Qualitative analysis:

Refers to visual assessment of the SI of the

lesions. The lesions SI were classified to low, intermediate and high signals compared to that of the myometrium.

B- Quantitative analysis:

ADC map were automatically generated by the

software on the basis of the images obtained. Mean

ADC values (multiplied by 10 -3mm2/s) were cal-

culated by drawing elliptical regions of interest

(ROI) in one representative region as large as

possible, however; care was taken to exclude

necrotic and cystic areas on the basis of findings

on T2-WI. We also avoid the normal junctional zone of the uterus, hemorrhagic areas and the

fibrous core of the polyps.

Nermin Y. Soliman, et al. 5145

Histopathologic correlation:

Radiological data of MRI were analyzed. MRI

data were compared to histopathological data; 82

patients underwent surgery, 9 patients underwent

exision of the lesion by hysteroscopy and 9 patients underwent biopsy after dilatation and curettage.

Statistical analysis: Statistical analysis was performed using statis-

tical package for the social science (SPSS) software

package version 22.0. Monte Carlo test was used

to compare SI of endometrial carcinoma and that

of benign diseases affecting the endometrium on

DWI. To compare the mean ADC values of benign

endometrial and subendometrial diseases one way analysis of variance (ANOVA) test was used and Student t-test was used to compare the mean ADC of endometrial carcinoma and that of benign dis-eases. Probability (p) values less than 0.05 were considered to be statistically significant. Receiver operating characteristic (ROC) analysis was used

to evaluate the diagnostic capability of mean ADC values for differentiation between endometrial

carcinoma and benign diseases. A value equal to the threshold or higher was considered to indicate a benign lesion, and a value below the threshold

was considered to indicate a malignant lesion.

Results

This study included 100 female patients with endometrial and subendometrial pathologies, their ages ranged from 20 to 85 years old.

According to the final histopathological diag-nosis, these patients were divided into two catego-ries: - Category I: Included 40 patients with endometrial

cancer. - Category II: Included 60 patients with benign

diseases affecting the endometrium. These were

subdivided into two groups of patients:

1- Group A: Included 41 patients with endome-trial diseases which were classified into (15 patients

with endometrial hyperplasia, 14 patients with

endometrial polyps, 6 patients with endometritis and 6 patients with molar pregnancy).

2- Group B: Included 19 patients with suben-dometrial diseases which were classified into (1 1 patients with ordinary leiomyomas, 3 patients with cellular leiomyomas and 5 patients with focal adenomyomas).

The most affected age group in patients with

endometrial cancer was 60-70 years in 18/40 pa-tients, as regard benign endometrial and suben-

dometrial diseases the most affected age group was 40-50 years in 20/60 patients. The mean age ± standard deviation of 40 patients with endometrial cancer (60.23 ±8.07) was significantly higher than that of 60 patients with benign diseases affecting the endometrium (48.67 ± 14.65) with p-value <0.001.

Abnormal vaginal bleeding was the most com-mon clinical presentation among our patients, as

it was seen in 96/100 patients. Other clinical pres-entations include: Pelvic pain in 27/100 patients, pelvi abdominal enlargement in 18/100 patients, vaginal discharge in 3/100 patients, dysmenorrhea in 2/100 patients and infertility in 1/100 patients.

Category I: Endometrial cancer: The details of the histopathological findings

were summarized in (Table 1).

Table (1): The final pathological diagnosis among the 40 cases

with endometrial cancer.

Pathology Frequency

(n) Percentage

(%)

Histopathological subtype: - Endometrial carcinoma type I 34 85 -- Endometrial carcinoma type II 6 15

• Serous papillary carcinoma 2 5 • Clear cell carcinoma 1 2.5 • Carcinosarcoma 1 2.5 • Large cell neuroendocrine 1 2.5

carcinoma • Undifferentiated carcinoma 1 2.5

Histological grade: - Grade 1 10 25 - Grade 2 15 37.5 - Grade 3 15 37.5

FIGO staging: - IA 11 27.5 - IB 6 15 - IIIC 14 35 - IVA 2 5 - IVB 7 17.5

Conventional MRI findings:

As regard T1-WI: 37/40 cases showed isoin-tense SI and 3/40 cases showed heterogeneous SI.

While; in T2-WI: 37/40 cases showed hypointense SI and 3/40 cases showed heterogeneous SI, if

compared with normal endometrium. As regard

study of early post contrast enhancement pattern:

33/40 cases showed weak enhancement, 5/40 cases

showed moderate enhancement and 2/40 cases

showed strong enhancement if compared to normal

myometrial enhancement.

As regard myometrial invasion: 25/40 cases

showed deep myometrial invasion, 9/40 cases


showed superficial myometrial invasion and 6/40

showed intact endometrial-myometrial interface.

D WI-MRI findings:

All studied cases with endometrial cancer dis-played high signal at DW images at (high b value

1000s/mm2) corresponding to low ADC values with restricted pattern at the ADC map. The mean

ADC value was found about 0.73 ±0.1 X 10-3

mm2/s.

Category II: Benign diseases affecting the en-dometrium: 1- Group (A): Endometrial diseases: a- Endometrial hyperplasia:

The final histopathological results were simple

hyperplasia without atypia in 12/15 patients and

complex hyperplasia without atypia in 3/15 patients.


As regard T1-WI: 14/15 cases showed isoin-tense SI and 1/15 cases showed hypointense SI.

While; in T2-WI: 8/15 cases showed heterogeneous SI, 5/15 cases showed isointense SI and 2/15 cases

showed hypointense SI if compared with normal endometrium. As regard study of early post contrast

enhancement pattern: 13/15 cases showed moderate

enhancement and 2/15 cases showed strong en-hancement if compared to normal myometrial

enhancement.

Junctional zone is intact in 13 /15 cases, how-ever; it is interrupted in 2/15 cases due to coexist-ence of adenomyosis.

D WI-MRI findings:

Endometrial hyperplasia displayed high signal at DW images at (high b value 1000 s/mm 2) which correspond to high ADC values with non restricted

pattern at the ADC map in 14/15 cases and low

signal at DW images corresponding to high ADC

values with free diffusion pattern at the ADC maps

in 1/15 case due to cystic changes. The mean ADC

value was found about 1.42 ±0.1 X10 -3mm2/s.

b- Endometrial polyp: Conventional MRI findings:

As regard T1-WI: 8/14 cases showed isointense SI and 6/14 cases showed heterogeneous SI. While;

in T2-WI: 7/14 cases showed heterogeneous SI, 6/14 cases showed hypointense SI and 1/14 case

showed isointense SI if compared with normal endometrium. As regard study of early post contrast

enhancement pattern: 8/14 cases showed strong

enhancement and 6/14 cases showed moderate

enhancement if compared to normal myometrial enhancement.

Specific findings were noted: T2-WI low SI fibrous core in 11/14 cases, T2-WI high SI intra-tumoral cysts in 10/14 cases and hemorrhage in

6/14 cases.

Endometrial-myometrial interface is intact in

11/14 cases, however; it is illdefined in 3/14 cases

due to coexistence of adenomyosis in one case and

large polypoidal mass thinning out the myometrium

in two cases.

DWI-MRI findings:

Endometrial polyps displayed low signal at DW images at (high b value 1000s/mm2) which correspond to high ADC values with free diffusion

pattern at the ADC map in 7/14 cases and interme-diate signal at DW images corresponding to high

ADC values with non restricted pattern at the ADC

map in 7/14 cases. The mean ADC value was found

about 1.51 ±0.10 X10 -3mm2/s.

c- Endometritis:

Conventional MRI findings: As regard T1-WI: 3/6 cases showed isointense

SI and 3/6 cases showed heterogeneous SI. While; in T2-WI: 3/6 cases showed heterogeneous SI, 3/6

cases showed hypointense SI if compared with

normal endometrium. As regard study of early post contrast enhancement pattern: 6/6 cases showed

peripheral strong enhancement if compared to

normal myometrial enhancement.

Concurrent pathologies were seen: Pelvic in-flammatory disease in 3/6 cases, retained product

of conception in 2/6 cases cancer cervix in 1/6 case.

All studied cases showed intact endometrial-myometrial interface.

DWI-MRI findings:

All studied cases with endometritis displayed high signal at DW images at (high b value 1000s/ mm2) corresponding to low ADC values with re-stricted pattern at the ADC map. The mean ADC value was found about 0.82 ±0.09 X10 -3mm2/s.

d- Gestational trophoplastic disease:

The final histopathological results were non

invasive complete mole in 3/6 patients, invasive complete mole in 2/6 patients and non invasive

partial mole in 1/6 patient.

Conventional MRI findings: As regard T1-WI: 6/6 cases showed heteroge-

neous low SI. While; in T2-WI: 6/6 cases showed heterogeneous high SI if compared with normal


endometrium. As regard study of early post contrast

enhancement pattern: 5/6 showed heterogeneous strong enhancement and 1/6 showed heterogeneous

moderate enhancement if compared to normal myometrial enhancement.

All studied cases showed distended uterus with characteristic heterogeneous snow storm appear-ance cystic like spaces with loss of uterine zonal

anatomy.

Specific findings were noted: Focal myometrial

contractions in 4/6 cases, hemorrhage in 6/6 cases,

intramyometrial vessels in 6/6 cases and theca

lutein cysts in 3/6 cases.

DWI-MRI findings:

All studied cases with molar pregnancy dis-played low signal at DW images at (high b value 1000s/mm2) corresponding to high ADC values with free diffusion pattern at the ADC map. The mean ADC value was found about 1.75 ±0.05 X10- 3mm2/s.

2- Group (B): Subendometrial diseases:

a- Ordinary leiomyoma: Conventional MRI findings:

As regard T1-WI: 8/11 cases showed isointense SI and 2/11 showed heterogeneous SI and 1/11 showed hyperintense SI. While; in T2-WI: 7/11

cases showed hypointense SI, 2/11 cases showed

isointense SI and 2/11 cases showed heterogeneous

SI if compared with normal myometrium. As regard study of early post contrast enhancement pattern:

9/11 showed weak enhancement and 2/11 showed

no enhancement if compared to normal myometrial

enhancement.

DWI-MRI findings:

Ordinary leiomyomas displayed low in 6/11 to intermediate signals in 5/11 cases at DW images

at (high b value 1000s/mm2) showing no significant reduction in ADC values with relative non restricted

pattern at the ADC map. The mean ADC value was

found about 1.37±0.11 X10 -3mm2/s.

b- Cellular leiomyoma:


As regard T1-WI: 3/3 cases showed isointense

SI. While; in T2-WI: 2/3 cases showed hyperintense

SI, 1/3 cases showed isointense if compared with normal myometrium. As regard study of early post contrast enhancement pattern: 3/3 showed moderate

enhancement.

DWI-MRI findings:

All studied cases with cellular leiomyomas displayed high signal at DW images at (high b value 1000s/mm2) corresponding to low ADC values with restricted pattern at the ADC map. The

mean ADC value was found about 0.9 ± 0.05 X 10-3mm2/s.

c- Focal adenomyoma: Conventional MRI findings:

As regard T1-WI: 3/5 cases showed isointense

SI and 2/5 cases showed heterogeneous SI. While; in T2-WI: 2/5 cases showed hypointense SI, 2/5 cases showed heterogeneous SI and 1/5 case showed isointense SI if compared with normal myometrium. As regard study of early post contrast

enhancement pattern: 4/5 cases showed moderate

enhancement and 1/5 case showed weak enhance-ment.

DWI-MRI findings:

All studied cases with focal adenomyomas

displayed intermediate signal at DW images at

(high b value 1000s/mm2) showing no prominent decrease of ADC values with relative non restricted

pattern at the ADC map. The mean ADC value was

found about 1.2±0.02 X10 -3mm2/s.

Table (2): Comparison between the mean ADC value among the 60 studied cases with benign diseases affecting

the endometrium.

Benign diseases ADC X10

-3mm

2/s

Mean ± SD Test of

significance

Endometrial hyperplasia 1.416±0.1 a F=44.4

Endometrial Polyp 1.5121 ±0.10 p<0.001* Endometritis 0.82±0.08b

Molar pregnancy 1.75±0.05 Ordinary leiomyoma 1.37±0.11 a

Cellular leiomyoma 0.90±0.06b

Focal adenomyoma 1.2±0.02

Total 1.36±0.32

F: One Way ANOVA test. *Statistically significant (p<0.05). - Similar superscripted letters in the same column denote non signif-

icant statistical differences between groups.

The overall difference (p<0.001) and the dif-ference between each group of benign endometrial and subendometrial pathologies mean ADC values was significant, however; the difference between

the mean ADC values of endometrial hyperplasia

and ordinary leiomyomas (p=0.58) as well as that of endometritis and cellular leiomyomas were

insignificant (p=0.29).

Endometrial

Benign Test of cancer diseases significance

DWI_SI (at high b value

1000 s/mm 2): Low:

n %

Intermediate: n %

High: n %

0 0%

0 0%

40 100%

21 35%

16 26.7%

23 38.3%

MC p<0.001*

MC: Monte-Carlo test. * Statistically significant (p<0.05).

All studied patients with endometrial cancer had a tendency to demonstrate high signal on DWI

at high b value 1000s/mm2, while benign diseases

Benign lesion Malignant lesion

Med

ian

AD

C *

10-

3mm

2

0.75

0.50

1.75

1.50

1.25

1.00

Specificity PPV NPV Accuracy AUC (95% CI)

0.97 (0.95-1.0)

Sensitivity (%)

97.5%

(%) (%) (%) (%)

88.3% 84.8% 98.1% 92.0%

Cut off point

ADC value X10 -3 mm2/s

≤0.915


Table (3): Comparison of DWI SI (at high b value 1000s/mm2) between 40 studied patients with endometrial cancer and 60 studied group with benign diseases.

showed variable appearance including low, inter-mediate and high signals with high significant difference (p<0.001).

Table (4): Comparison of the mean ADC value between 40

studied patients with endometrial cancer and 60 studied patients with benign diseases.

Endometrial cancer N=40

ADCx10-3 mm2/s Mean ± SD 0.73±0.10 1.36±0.32 t=13.87 Median 0.73 1.39 p<0.001* (range) (0.55-1.0) (0.68-1.8)

t: Student t test. * Statistically significant (p<0.05).

The mean ADC value of endometrial cancer

was 0.73±0.10 x10 -3mm2/s and for benign diseases, it was 1.36±0.32 x10-3mm2/s which was signifi-cantly higher than that of endometrial cancer with

p-value <0.001 (Fig. 1).

Benign diseases N=60

Test of significance

Fig. (1): Box and Whisker plots demonstrating ADC values of endometrial cancer and benign endome-trial diseases. The center horizontal line indicates the median value. The bottom and top edges of the box

indicate the 25 th and 75 th percentiles. The vertical line indicates the range of data.

Table (5): The valididity of mean ADC value in differentiation between endometrial cancer and benign diseases

affecting the endometrium.

AUC: Area Under Curve. PPV: Positive predictive value. NPV: Negative predictive value.

Sen

siti

vity

1.0

0.8

0.6

0.4

0.2

0.0

Fig. (2): ROC curve showing the validity of mean ADC in differentiation between endometrial cancer

and benign diseases. The area under the (ROC) curve

was 0.97. A cut off value of 0.915 X10 -3mm2/s was considered to be the threshold. When an ADC value of 0.97 X10 -3mm2/s or more was estimated the lesion was considered to be benign and a value below the threshold was considered to indicate endometrial

cancer with the sensitivity, specificity and accuracy of 97.5%, 88.3% and 92.0% respectively.

0.0 0.2 0.4 0.6 0.8 1.0

1-Specificity Diagonal segments are produced by ties


Case (1): Pathologically proven grade 3 endometrioid adenocarcinona. Figs. (A) Sagittal. (B) Axial T2-WI showed low SI endometrial

mass and bilateral enlarged pelvic lymph nodes (white arrows). (C) Axial. (D) Sagittal post contrast T1 fat suppressed images demonstrated

weak enhancement of the lesion and heterogeneous enhancement of regional and retroperitoneal lymph nodes (white arrows). (E) Axial, (G)

sagittal DWI (b=1000 s/mm2) showed deep myometrial invasion (black arrow) (F) axial, (H) sagittal ADC map showed restricted diffusion

pattern of the mass. The ADC value was measured=0.63 x10 -3 mm2/s, bilateral enlarged local pelvic lymph nodes in (E,F) and retroperitoneal enlarged lymph node in (G,H) displaying high SI at DWI and low SI at ADC maps similar to the primary tumor (white arrows).


Case (2): Pathologically proven simple endometrial hyperplasia without atypia. Figs. (A) Sagittal (B) Axial T2-WIs showed diffuse

endometrial thickening displaying heterogeneous SI with picture of adenomyotic changes in the form of irregular endometrial-myometrial

junction, thickened junctional zone and high SI linear striation on T2-WI extending from the endometrium into the myometrium with subsequental

pseudowidening of the endometrium. (C) Sagittal post contrast T1-W fat suppressed image shows global enhancement following the normal endometrium. (D) Sagittal DWI ( b=1000 s/mm

2) shows high SI within the endometrium, with corresponding high SI at (E) Sagittal ADC map

denoting non restricted diffusion pattern. The ADC value was measured=1.44 x10-3

mm2

/s.

Case (3): Pathologically proven endometrial polyp. Figs. (A) Sagittal T2-WI shows focal endometrial thickening displaying heterogeneous

SI with hypointense linear SI component suggesting fibrous core (white arrow) at (B) axial T2-WI with intact endometrial junctional zone interface. (C) Sagittal post contrast T1-W fat suppressed image shows avid strong partial enhancement. Axial DWI ( b= 1000 s/mm

2) the lesion

displaying intermediate SI, corresponding to high SI at (E) Axial ADC map denoting non restricted diffusion pattern. The ADC value was found

1.52 x10-3

mm2/s.N.B: [Low SI fibrous core on DWI and ADC map (black arrow), also small intratumoral cyst displaying T2 shine through

effect in the form high SI on both DWI and ADC map (black asterisk) are noted].


Case (4): Pathologically proven submucosal uterine fibroid. Figs. (A) Sagittal T2-WI (B) axial T2-WI showing a well defined isointense

mass compared to the myometrium involving the lower part of the endometrial cavity, the related posterior wall of the myometrium and extending

into the endocervical canal through a dilated internal os.(C) Sagittal post contrast T1-W fat suppressed image showed weak enhancement of

the mass. (D) Axial DWI ( b=1000 s/mm2) shows the mass displaying intermediate SI, with no significant reduction of SI at (E) Axial ADC map denoting non restriced diffusion pattern. The ADC value was measured=1.30 x10

-3 mm2/s.


Case (5): Pathologically proven high grade non endometrioid adenocarcinoma (small cell neuroendocrine carcinoma). Figs.

(A) Sagittal T2-WI demonstrated bulky uterus with an endometrial polypoidal mass of heterogeneous SI (B) Axial T2-WI showed

bilateral enlarged pelvic lymph nodes. (C) Sagittal T1-W fat suppressed image shows heterogeneous enhancement. (D) Sagittal

DWI (b=1000 s/mm2) (E) Sagittal ADC map showed infilterating tumor diffusely involving the myometrium, the cervix and

the vagina with restricted diffusion pattern. The ADC value was measured= 0.55 x10 -3 mm2/s. (F) Axial DWI bilateral enlarged pelvic lymph nodes (black arrows) showed high SI and low SI at (G) axial ADC map similar to the primary tumor.

Discussion

We investigated the diagnostic efficacy of DWI

and ADC values in 100 female patients with en-dometrial and subendometrial pathologies, their ages ranged from 20-85 years, presented with abnormal vaginal bleeding in (96%) of studied cases and how can DWI and the mean ADC values

provide a diagnostic differentiation between en-dometrial carcinoma and benign endometrial dis-eases.

Endometrial cancer occurs most frequently in

women during the sixth and seventh decades of

life [7] , our study revealed statistically significant

difference between the age group having endome-trial cancer (mean age ± SD 60.23±8.07) and those having benign endometrial diseases (mean age ± SD 48.67± 14.65), so endometrial cancer affected

older age group. This was in agreement with study conducted by Bakir et al., [10] .

In this current study, morphological MRI signal

intensities were not significantly different in dis-crimination between malignant and benign diseases affecting the endometrium this was matched with

the results of the study performed by Kierans et

al., [11] . We found different pathological types of

endometrial and subendometrial diseases with variable appearance on DWI and ADC map.

Our study revealed that all studied cases with endometrial carcinoma were more likely to have

restricted diffusion due to high cellularity in the form of high signal on high b-value (1000s/mm2) DWI and low ADC value (0.73 ±0.1 x10 -3mm2/s) this was almost similar to the results of study

performed by Shady, et al., [12] .

Addley et al., [13] stated that not all lesions with restricted diffusion on DWI are malignant, and there are many benign diseases can exhibit restricted diffusion pattern, this clarifies the im-portance of the use of DWI as an adjunct rather than a replacement for morhological imaging se-quences in lesion characterization. We had (9)

pathologically proved benign diseases with restrict-ed diffusion pattern included (3) cellular leiomyo-mas and (6) endometritis.

According to Duarte et al., [14] the presence of restricted diffusion within cellular leiomyomas due to their characteristic high cellularity, so image

analysis should always be made in correlation with

the appearances on T1 and T2 WI. In endometritis,

the presence of pus which is a fluid with high

viscosity and dense cellularity can explain the high

SI in DWI and the low ADC value, after contrast

study, the pus shows abscent enhancement, which

can be easily differentiated from malignant tumors

with enhanced solid parts.

In DW imaging qualitative analysis, we found that the majority of endometrial hyperplasia exhib-ited high SI and only one case displayed low SI

due to cystic changes secondary to tamoxifen

therapy this was matched with Bakir et al., [10] and Ahmed et al., [15] who observed that qualitative analysis of diffusion image alone cannot distinguish

endometrial lesions so quantitative analysis with

the ADC map is fundamental for endometrial le-sions characterization. We found that mean ADC

value for endometrial hyperplasia was (1.42 ±0.1 x10 -3mm2

/s) this was in accordance with results

of study performed by Elsammak et al., [2] . The high ADC value observed in endometrial hyperpla-sia can be explained by prominant extracellular


area with edematous tissue and cystic structures

[16] .

Our study showed that there was high frequency

of low SI T2-WI fibrous core and high SI T2-WI

intratumoral cystic changes that can be seen by

conventional MRI in endometrial polyps which reflect their internal structure, as also stated with

Wang et al., [3] .

We found in the current study that the signal

intensity of endometrial polyps in qualitative anal-ysis of DW imaging was intermediate or low SI this was matched with results obtained by Wang et al., [1] and Ahmed et al., [15] studies. We found also that the mean ADC value was (1.51 ±0.10 x 10-3mm2/s) this was in accordance with results of

study performed by Bakir et al., [10] who explained the high ADC value in endometrial polyps due to their histopathologic composition of dense fibrous

stroma or smooth muscle, with cystic hyperplasia

of endometrial glands, may cause decreased cellular

density.

In our study, we found that there was a signif-icant statistical difference between mean ADC

value between endometrial polyps and endometrial

hyperplasia, in contrast to the results of the study

performed by Elsammak et al., [2] . It may be due to avoidance of hemorrhagic regions and the fibrous core in the current study, instead of selecting the

largest ROI within the lesion in ADC map.

The current study revealed that the signal in-tensity of submucosal ordinary leiomyomas in

qualitative analysis of DW imaging was interme-diate or low SI this was matched with Ahmed et

al., [15] & Sato et al., [17] studies and we found that the mean ADC value was (1.37 ±0.10 x10 -3

mm2/s) this was in accordance with results of study

performed by Elsammak et al., [2] .

We also found that the mean ADC value for endometrial polyps was significantly higher than that for submucosal ordinary leiomyomas this in agreement with Karakas, et al., [9] who stated that leiomyomas can show variable histopathologic features in association with degeneration such as

hyalinization and edema. Hyaline changes lead to homogenous eosinophilic strips accumulation in the extracellular area, and the extracellular space

narrows. Thus, this histological feature decreases ADC values; however, edema causes increased

ADC values due to extracellular space widening.

As regard cellular leiomyomas, we found that its mean ADC value was (0.9 ±0.05 x10 -3mm2/s), Arlea et al., [18] reported that a value less than 1.1

categorized the lesion in a high-risk group, an ADC value of 1.1 or higher put the lesion as low-risk

group.

Although MRI is often able to differentiate between submucosal endometrial lesions including

focal adenomyoma and leiomyomas, sometimes their imaging features may overlap [19] . In the qualitative analysis of DW imaging, we found that

focal adenomyoma exhibited intermediate SI this was matched with study performed by Takeuchi

and Matsuzaki, [20] . We found that the mean ADC value was (1.2±0.02 x10-3mm2/s) this was in ac-cordance with results of study performed by Kilick-esmez et al., [21] . We found also that the mean ADC value is significantly lower than fibroids

similar to the results of previous studies performed

by Tian et al., [22] , Jha et al., [19] and Kilickesmez et al., [21] .

Another benign endometrial disease that we

have investigated in our study is gestational tro-phoplastic disease (GTD). Lima et al., [23] stated that diagnosis of GTD should depend on a clinical history, physical examination, serum B-HCG level and pelvic ultrasound. The evaluation of GTD does

not include MRI routinely, which is recommended in complicated and suspicious cases. We found in our study that GTD had specific radiological find-ings in conventional MRI included snow storm appearance, tortuous myometrial vessels, focal

hemorrhagic areas and myometrial contraction this

was in accordance with Sefidbakht et al., [24] . By DWI qualitative analysis all cases displayed low SI with mean ADC value 1.75±0.05 which was significantly higher than the other benign endome-trial diseases in the current study being vesicular

and cystic, if compared to most other tumors, rather

than cellular and dense this was in accordance with results of study performed by Sefidbakht et al.,

[24] .

Haldorsen and Salvesen, [25] stated that the mean ADC values of endometrial cancer have been

reported to range between (0.75-0.97 x10 -3mm2/s) was statistically significant lower than that reported for benign endometrial lesions (1.21-1.76 x10 -3

mm2/s) in different studies.

In our study, a significant difference (p=<0.001) was found between the mean ADC values of en-dometrial cancer (0.73 ±0.1 x10-3mm2/s) and benign endometrial diseases (1.36 ±0.32 x10 -3 mm2/s). Using (0.915 x10 -3mm2/s) as a cut off value for distinguishing between endometrial cancer and

benign endometrial diseases had 97.5% sensitivity,

88.3% specificity and 92.0% accuracy.


Kececi et al., [26] reported that the mean ADC value of the endometrial cancer (0.94 ±0.18 x10 -3

mm2/s) was statistically significantly lower than the ADC value of benign lesions (1.45 ±0.22 x10

–3

mm2/s) and the ADC threshold value to determine

whether the lesions were benign or malignant was

(1.10 x10 -3mm2/s) with 85.7% sensitivity and 92.8% specificity.

Elsammak et al., [2] ; study revealed also that a significant difference was found between the mean

ADC values of malignant masses (0.82 x10 -3

mm2/s) and benign lesions (1.44 x10 -3mm2/s), they had 88.9% sensitivity, 100% specificity and accu-racy 88.9% when used (1.19 x10 -3mm2/s) as a cut off value for differentiation benign from malignant

endometrial lesions.

In a related context Keriakos and Darwish, [27] & Cavusoglu et al., [16] reported that the mean ADC value was (0.8 x10 -3mm2/s) for endometrial cancer with best cut off value of ADC to discrim-inate between benign and malignant lesions was (1.2 x10 -3mm2/s).

Wang et al., [3] reported that various cut off ADC values have been used to differentiate malig-nant from benign endometrial lesions. These cut

off ADC values ranged from (0.908 x10 -3mm2/s) to (1.28 x10 -3mm2/s). By applying the cut off ADC

value of (0.98 x10 -3mm2/s) in their study, they achieved 100% sensitivity, 97.1% specificity and accuracy 98.3%.

The results of the previous studies were slightly

higher than our results, According to Kececi et al.,

[26] and Wang et al., [3] the differences in ADC values in the studies can be explained by the dif-ference in many technical variables that affect the

ADC value such as different MRI units, pulse sequences, or b values.

Both in the light of the literature, as well as the

results of our study, it can be said that DWI and ADC measurements have a high diagnostic efficacy

in discrimination of endometrial cancer from benign

diseases affecting the endometrium, since DWI is a non time consuming and safe technique without ionizing radiation that is not need contrast material,

thus enabling its use in patients with impaired renal

function, it can provide qualitative and quantitative valuable information to morphological MR se-quences and can easily be added into standard imaging protocols using MRI.

Limitations & recommendations: The relatively small number of patients as the

study was limited only to patients with histopatho-

logically diagnostic lesions in order to obtain

undebatable reference standards so further studies

on a larger population are required to confirm our results.

DWI in the pelvis is still limited by technical problems that can affect image quality, as motion

artifact additionally the low resolution of DWI

that become more worse at higher b value. Fused T2-WI and high b value imaging with background body signal suppression (DWIBS) can improve anatomic data.

ADC measurements can be affected by the site, size, and form of the ROI. The circular ROIs placed

in the selective region of the lesion were applied

in our study. The selection of the area in the lesions can be subjective and might lead to some bias. Histogram analysis of the entire tumor may be a

more accurate method to avoid sample selection

bias.

References

1- WANG J., YU T., BAI R., SUM H., ZHAO X. and LI Y.: The value of the apparent diffusion coefficient in differ-entiating stage IA endometrial carcinoma from normal

endometrium and benign diseases of the endometrium: Initial study at 3-T magnetic resonance scanner. J. Comput Assist Tomogr., 34 (3): 332-337, 2010.

2- ELSAMMAK A., SHEHATA S.M., ABULEZZ M. and GOUHAR G.: Efficiency of diffusion weighted magnetic

resonance in differentiation between benign and malignant

endometrial lesions. The Egyptian Journal of Radiology and Nuclear Medicine, 48 (3): 751-759, 2017.

3- WANG X., ZHAO Y., HU Y., ZHOU Y., YE X., LIU K., et al.: Evaluation and validation of the diagnostic value

of the apparent diffusion coefficient for differentiating

early-stage endometrial carcinomas from benign mimickers

at 3T MRI. Oncotarget, 8 (28): 46390-46397, 2017.

4- RECHICHI G., GALIMBERTI S., SIGNORELLI M., FRANZESI C.T., PEREGO P., VALSECCHI M.G., et al.: Endometrial cancer: correlation of apparent diffusion

coefficient with tumor grade, depth of myometrial inva-sion, and presence of lymph node metastasis. AJR, 197: 256-262, 2011.

5- MOTOSHIMA S., IRIE H., NAKAZONO T., KAMURA T. and KUDO S. Diffusion-weighted MR imaging in gynecologic cancers. J. Gynecol. Oncol., 22 (4): 275-287,

2011.

6- INOUE C., FUJII S., KANEDA S., FUKUNAGA T., KAMINOU T., KIGAWA J., et al. Apparent Diffusion Coefficient (ADC) Measurement in Endometrial Carci-noma: Effect of Region Interest Methods on ADC Values.

Journal of Magnetic Resonance Imaging, 40: 157-161,

2014.

7- HABIB L.A., ALHAWARY M.M., ABD ELDAYEM E.H. and ABD ALGHANY A.F.: Role of MRI in Diagnosis of Endometrial Cancer. The Egyptian Journal of Hospital

Medicine, 72 (10): 5505-5512, 2018.


8- VERMOOLEN M.A., KWEE T.C. and NIEVELSTEIN R.A.J.: Apparent diffusion coefficient measurements in the differentiation between benign and malignant lesions: A systematic review. Insights Imaging, 3 (4): 395-409, 2012.

9- KARAKAS O., KARAKAS E., DOGAN F., KILICAS-LAN N., CAMUZCUOGLU A., INCEBIYIK A., et al.: Diffusion-weighted MRI in the differential diagnosis of uterine endometrial cavity tumors. Wien Klin Wochenscr,

127 (7-8): 266-273, 2015.

10- BAKIR V.L., BAKIR B., SANLI S., YILDIZ S.O., LYI-BOZKURT A.C., KARTAL M.G., et al.: Role of diffusion-weighted MRI in the differential diagnosis of endometrioid

and non-endometrioid cancer of the uterus. Acta. Radiol., 58 (6): 758-767, 2017.

11- KIERANS A.S., BENNETT G.L., HAGHIGHI M. and ROSENKRANTZ A.B.: Utility of conventional and dif-fusion-weighted MRI features in distinguishing benign from malignant endometrial lesions. Eur. J. Radiol., 83

(4): 726-732, 2014.

12- SHADY M.S., BAKRY M.A., MAZROA J.A. and GADELHAK B.N.: MR diffusion imaging in preoperative

evaluation of depth of myometrial invasion in endometrial

carcinoma. The Egyptian Journal of Radiology and Nuclear

Medicine, 47: 611-619, 2016.

13- ADDLEY H., MOYLE P. and FREEMAN S.: Diffusion-weighted imaging in gynaecological malignancy. Clin. Radiol., 72 (11): 981-990, 2017.

14- DUARTE A.L., DIAS J.L. and CUNHA T.M.: Pitfalls of diffusion-weighted imaging of the female pelvis. Radiol.

Bras., 51 (1): 37-44, 2018.

15- AHMED S.A., EL TAIEB H.A. and ABOTALEB H.: Diagnostic performance of sonohysterography and MRI diffusion in benign endometrial lesion characterization. The Egyptian Journal of Radiology and Nuclear Medicine, 49: 579-589, 2018.

16- CAVUSOGLU M., CILIZ D.S., OZSOY A., DURAN S., ELVERICI E., ATALAY C.R., et al.: Diffusion-Weighted MRI of Postmenopausal Women wih Vaginal Bleeding and Endometrial Thickening: Differentiation of Benign and Malignant Lesions. Belgian Society of Radiology, 70: 1-10, 2016.

17- SATO K., YUASA N., FUJITA M. and FUKUSHIMA Y.: Clinical application of diffusion-weighted imaging for preoperative differentiation between uterine leiomyoma

and leiomyosarcoma. Am. J. Obstet. Gynecol., 210 (4): 368.e1-368.e8, 2014.

18-ARLEO E.K., SCHWARTZ P.E., HUI P. and McCARTHY S.: Review of Leiomyoma Variants. AJR, 205: 912-921, 2015.

19- JHA R.C., ZANELLO P.A., ASCHER S.M. and RAJAN

S.: Diffusion-weighted imaging (DWI) of adenomyosis and fibroids of the uterus. Abdom Imaging, 39: 562-569,

2014.

20- TAKEUCHI M. and MATSUZAKI K.: Adenomyosis: Usual and Unusual Imaging Manifestations, pitfalls, and

problem-solving MR Imaging Techniques. RadioGraphics, 31: 99-115, 2011.

21- KILICKESMEZ O., BAYRAMOGLU S., INCI E., CIMIL-LI T. and KAYHAN A.: Quantitative diffusion-magnetic resonance imaging of normal and diseased uterine zones. Acta. Radiol., 50 (3): 340-7, 2009.

22- TIAN T., ZHANG G.F., ZHANG H. and LIU H.: Intravoxel incoherent motion diffusion-weighted imaging in differ-entiating uterine fibroid from focal adenomyosis: Initial

results. SpringerPlus, 5: 9, 2016.

23- LIMA L.L.A., PARENTE R.C.M., MAESTA I., AMIM JUNIOR J., REZENDE FILHO J.F., MONTENEGRO C.A.B., et al.: Clinical and radiological correlations in

patients with gestational trophoblastic disease. Radiol. Bras., 49 (4): 241-250, 2016.

24- SEFIDBAKHT S., HOSSEINI F., BIJAN B., HAMEDI B. and AZIZI T.: Qualitative and Quantitative Analysis of Diffusion-Weighted Imaging of Gestational Trophob-lastic Disease: Can It Predict Progression of Molar Preg-nancy to Persistent Form of Disease? European Journal

of Radiology, 88: 71-76, 2017.

25- HALDORSEN I.S. and SALVESEN H.B.: What Is the Best Preoperative Imaging for Endometrial Cancer? Curr.

Oncol. Rep., 18 (4): 25, 2016.

26- KECECI I.S., NURAL M.S., ASLAN K., DANACI M., KEFELI M. and TOSUN M.: Efficacy of diffusion-weighted magnetic resonance imaging in the diagnosis

and staging of endometrial tumors. Diagn. Interv. Imaging,

97 (2): 177-86, 2016.

27- KERIAKOS N.N. and DARWISH E.: Diffusion weighted imaging in suspicious uterine tumors; how efficient is it?

The Egyptian Journal of Radiology and Nuclear Medicine, 49 (3): 838-845, 2018.

role of diffusion-weighted magnetic resonance imaging in

Documents